基于三维超声构建糖尿病肾病诊断预测模型的初步研究

Development of a predictive model for diagnosis of diabetic nephropathy based on three-dimensional ultrasound

目的应用三维超声获得肾脏体积参数,结合临床资料指标建立诊断糖尿病肾病的预测模型方程。方法选取2型糖尿病合并肾损害的研究对象62例,根据肾脏穿刺活检病理结果将其分为糖尿病肾病(DN)组和非糖尿病肾病(NDRD)组,其中DN组35例,NDRD组27例,通过三维超声得出肾脏体积,并用其体表面积进行校正得出肾脏体积指数,比较分析两组研究对象的性别、年龄、体重、身高、体重指数、体表面积、收缩压/舒张压、尿蛋白、肾小球滤过率、血浆尿素氮、血肌酐、尿肌酐、空腹血糖、糖尿病病史、是否进行糖尿病治疗、是否有糖尿病视网膜病变、是否有血尿以及双肾体积、肾脏体积指数是否存在差异,选取临床资料指标与肾脏体积指数用Logistic回归进行预测模型的建立。结果DN组与NDRD组比较,收缩压、尿蛋白、肾小球滤过率、血浆尿素氮、血肌酐、糖尿病病史、是否有糖尿病视网膜病变、是否有血尿临床资料,差异均有统计学意义(t=4.8056、2.3748、5.0350、4.0205、4.3821、5.9283,χ^2=2.9606、3.1691,P均<0.05),右肾肾脏体积指数两组间的比较,差异有统计学意义(t=2.7166,P<0.05),性别、年龄、体重、身高、体重指数、体表面积、舒张压、尿肌酐、空腹血糖、是否进行糖尿病治疗、右肾体积、左肾体积和左肾体积指数,差异均无统计学意义(P均>0.05);通过双肾肾脏体积指数和主要临床资料参数构建诊断预测模型方程,预测模型方程判断DN的ROC曲线下面积为0.9217,95%置信区间(0.8557~0.9877),最佳阈值0.2069,特异度85.19%,敏感度85.71%,准确性85.48%,阳性预测值0.8824,阴性预测值0.8214,阳性似然比5.7857,阴性似然比0.1677。结论基于三维超声建立的DN诊断的预测模型方程,可为临床诊断DN提供了一种新的诊断手段,对今后的临床诊断和治疗有重要的意义。

Objective To obtain renal volume parameters by using three-dimensional ultrasound,and to establish a predictive model for the diagnosis of diabetic nephropathy(DN)based on clinical data.Methods Sixty-two subjects with type 2 diabetes mellitus complicated with renal damage were enrolled.Based on the pathological results of renal biopsy,the patients were divided into either a DN group(n=35)or a non-diabetic renal diseases(NDRD)group(n=27).Kidney volume was obtained by three-dimensional ultrasound,and the body volume index was used to obtain the kidney volume index.Gender,age,weight,height,body mass index,body surface area,systolic blood pressure/diastolic blood pressure,urinary protein,glomerular filtration rate,plasma urea nitrogen,serum creatinine,urinary creatinine,fasting blood glucose,history of diabetes,diabetes treatment,diabetic retinopathy,hematuria,kidney volume,and kidney volume index were compared between the two groups.Clinical data indexes and kidney volume index were selected by Logistic regression to establish a predictive model for the diagnosis of DN.Results Urinary protein,glomerular filtration rate,plasma urea nitrogen,serum creatinine,history of diabetes,diabetic retinopathy,and hematuria differed significantly between the two groups(t=4.8056,2.3748,5.0350,4.0205,4.3821,5.9283;χ^2=2.9606,3.1691;all P<0.05).There was also a statistically significant difference in the right kidney volume index between the two groups(t=2.7166,P<0.05).There was no significant difference in gender,age,weight,height,body mass index,body surface area,diastolic blood pressure,urinary creatinine,fasting blood glucose,diabetes treatment,volume of both kidneys,or left kidney volume index(all P>0.05).A predictive model was constructed by using renal kidney volume and the main clinical parameters.The area under the ROC curve of this model for diagnosing DN was 0.9217(95%confidence interval:0.8557-0.9877);with an optimal threshold of 0.2069,the specificity,sensitivity,accuracy,positive predictive value,negative predictive value,positive likelihood ratio,and negative likelihood ratio were 85.19%,85.71%,85.48%,0.8824,0.8214,5.7857,and 0.1677,respectively.Conclusions The predictive model for the diagnosis of DN based on three-dimensional ultrasound provides a new diagnostic method for DN,which is of great significance for the clinical diagnosis and treatment of DN.